Analytics in Industry 4.0: Investigating the Challenges of Unstructured Data

Digital Twins in Production: The Integration of Semi- and Unstructured Data

“…The broader the database, the more production related insights can be gained. It extends the current work of the data integration challenges of digital twins and the usage of unstructured data in Industry 4.0 [10,18] by answering how to integrate log files. Furthermore, the prototype can be used to improve the current production processes as well as issue handling to react fast, targeted and cost oriented.…”

“…The integration of different variants of data is a current challenge for implementing digital twins [10]. But, without the integration of also semi-and unstructured data like textual data (e.g., log files), images or sound into digital twins the intended big picture of the related physical object (e.g., machinery or production line) is not possible [18,17,10]. In general, companies mostly use within their digital twins structured data like temperature, pressure, time, etc.…”

“…The implementation can be done via the fastapi Python framework. Furthermore, more other unstructured data sources [18] like images, sound, etc. can be integrated via a REST-API into a digital twin.…”

“…As it shows, the production related digital twin is not only visualizing the current state of an e.g., machinery or good, based on the structured sensor data like temperature. It connects and integrates also to semi-and unstructured data [18,10] sources and enables a broader view and deeper investigations (e.g., for root cause analysis through process and text mining results). The digital twin's visualization can be realized in different ways like in a self-developed app, using a framework or standard data visualization solutions, or via dashboard tools like Tableau, QlikView, Grafana, Microsoft PowerBI, etc.…”